Installed msysGit, wrote the code to start ssh-agent in .profile, everything worked.
Installed cygwin, without Git, just ssh and cURL.
SSH Agent no longer starting when Git bash opens.
I can start a new ssh-agent process instance, I see it when running ps in the Git bash, but when trying to use ssh-add, I get this error:
Could not open a connection to your authentication agent.
With cygwin, lots of .profile and .bashrc files were created in it's install folder (C:\cygwin). Not sure if this is the issue.
How can I fix this, please?
Take a look at my answer posted here for the ssh-add issue. Hopefully, this solution should work in your scenario too.
Related
I haven't been able to find a similar question, but I can't imagine this is an issue that comes up too much.
At my workplace I've decided to switch over to a full Linux experience by using Cygwin, over what I was previously using, Git Bash.
The only slight drawback is that we have a script written in Ruby that deploys changes to our site. I can run this from my Git Bash terminal no problem, but not Cygwin as there doesn't seem to be a Ruby installation on here. I've learned about sending standard output to other terminal windows, but I've only been successful when both terminals have been the same shell.
What I'd like to do is the following:
1. Type a command into Cygwin that opens a Git Bash terminal and automatically runs ./deploy.sh on that Git Bash terminal upon opening.
2. If possible, also pass in an argument in the same command, that the script expects upon running. (In this case it's whether we want to deploy to our staging or production server).
Is this possible? So far I can get what looks to be a Git Bash terminal to open and prompt me for the server, but when I enter it, it does not work. I get an error with regards to a command within the script not being found. Thing is, running this same ./deploy.sh script in a Git Bash terminal that I open myself works perfectly as it always has. The Git Bash terminal that is being opened by Cygwin does not appear to have all of the capabilities of a Git Bash terminal I open 'the normal way'.
The command I am running so far from Cygwin is as follows:
/cygdrive/c/'Program Files'/Git/git-bash.exe "./deploy.sh"
I understand there is probably a way of getting this to work by installing Ruby via Cygwin but I'd like to mess around with my work environment as little as possible in case I somehow affect my ability to deploy my work... and besides, Git Bash definitely runs this script without an issue so I can't see why I would need a duplicate installation.
Many thanks!
I'm trying to add a build step in jenkins to copy files from my build server to my web application server. I've got the following command working in the command prompt
sudo scp -r /var/lib/jenkins/workspace/demoproj/publish root#0.0.0.0:/usr/temp
but when I run this command, it prompts me for a password every time. I found out about sshpass, but when I run this command...
sudo sshpass -p "passwordhere" scp -r /var/lib/jenkins/workspace/demoproj/pub root#0.0.0.0:/usr/temp
the terminal gets stuck. And never makes it through.
My main problem is if I add the first command to a build step in jenkins, it won't be able to pass the password over. How can I either supply the password in jenkins, or modify the command to pass over my credentials?
Helpful information: I'm using Putty on Windows 10 to connect to my
Ubuntu 16.04.3 LTS x64 servers from another Ubuntu 16.04.3 server.
First, sshpass needs to be installed on both the systems that is, the one running your jenkins instance as well as the one you are trying to access that is: root#0.0.0.0. You can verify it by doing 'which sshpass' or 'whereis sshpass'. If its not installed even in one of them then you need to install it first.
Also, Have you ever tried doing a ssh to the said machine: root#0.0.0.0 from the system where you have your jenkins instance? If not then there might not be an entry in the 'known-hosts' of either system. for that you can do ssh with '-o StrictHostKeyChecking=no' option to make an automatic entry in known-hosts.
Alternatively, if you dont want to enter password again and again you should work with 'keys'. Generate a unique key for both the systems and do an scp or ssh with -i option.
You should use jenkins credentials instead of using sensitive passwords directly into the scripts. Put the whole scp or ssh part inside a block which looks like: withCredentials(){}.
What's the point of having CI if you are required to be nearby to enter password every time? Install "publish over ssh" plugin, it has a step to send stuff over ssh.
https://wiki.jenkins.io/display/JENKINS/Publish+Over+SSH+Plugin
Look at "Use SSH during a build" section, you can use "send files or execute commands over SSH" build step. This shall become available after plugin installation.
I am having some difficulty running jobs over SSH. I have a series of networked machines which all have access to the same Home folder (when my executable is installed). While working on one machine I would like to be able run my code through ssh using the following sort of command:
ssh -q ExecutableDir/MyExecutable InputDir/MyInput
If I ssh in to any of the machines I wish to run the job on remotely and simply run:
ExecutableDir/MyExecutable InputDir/MyInput
It runs without fail, however when I run through SSH I get an error saying some shared libraries can't be found. Has anyone come across this sort of thing before?
ok I figured it out myself.
It seems when you run things through ssh in the way shown above you don't inherit the path variables etc. that you would if you ssh-ed in 'properly'. You can see this by running:
ssh RemoteMachine printenv
and comparing the output to what you would normally get if you were connected to the remote machine. The solution I then went for was to run something like the following:
ssh -q ExecutableDir/MyExecutable source ~/.bash_profile && InputDir/MyInput
Which then gets all the paths and stuff you might need from the bash_profile file on the remote machine
When I execute this command below:
scp -P 36000 hdfs#192.168.0.114:~/tmp.txt SOQ_log.txt
I get an error:
command-line: line 0: Bad configuration option: PermitLocalCommand
Does anyone know why?
scp runs a copy of the ssh program to create the communications channel, and it runs ssh with the options:
-oForwardAgent=no -oPermitLocalCommand=no -oClearAllForwardings=yes
So that explains where the "PermitLocalCommand" option is coming from in the first place. I'll add that sftp uses the same options to run ssh, so it'll probably display the same behavior.
"PermitLocalCommand" is normally a valid ssh configuration option. If your copy of ssh is complaining about it, then it seems that your copy of ssh isn't the normal copy of ssh that goes with your copy of scp.
This serverfault question suggests that the error could be due to someone installing a malware version of ssh (ie, a rootkit) on your system. This forum thread also suggests that the problem is due to having an altered version of ssh, which was fixed by removing and reinstalling the OpenSSH client utilities.
An alternate explanation would be that someone--maybe your Linux distro maintainer--has installed a version of ssh on your system with that option removed, and you're using it unawares. Or you have a very old version of the ssh program for some reason, which doesn't support the option.
My system is CentOs 5.9
I'm facing the same problem, I found it to be due to this configuration line in /etc/ssh/sshd_config:
# override default of no subsystems
Subsystem sftp /opt/libexec/sftp-server
But I cannot run /opt/libexec/sftp-server, it is broken for some reason
now it is solved by reinstall the remote openssh-server:
yum erase openssh-server
yum install openssh-server
now the changes to
# override default of no subsystems
Subsystem sftp /usr/libexec/openssh/sftp-server
and /usr/libexec/openssh/sftp-server is runnable
don't forget to execute:
/etc/init.d/sshd restart
Sometimes command cannot parse this kind of stuff
:~/
Id change it to the full path.
Somehow, I can't run vagrant or heroku in cygwin. It works fine when I'm using the default windows cmd application, but in cygwin, I get this error for vagrant:
C:/vagrant/vagrant/embedded/lib/ruby/site_ruby/1.9.1/rubygems/dependency.rb:247:in to_specs': Could not find vagrant (>= 0) amongst [] (Gem::LoadError) from C:/vagrant/vagrant/embedded/lib/ruby/site_ruby/1.9.1/rubygems/dependency.rb:256:into_spec' from C:/vagrant/vagrant/embedded/lib/ruby/site_ruby/1.9.1/rubygems.rb:1231:in gem' from C:/vagrant/vagrant/embedded/gems/bin/vagrant:22:in'
And for heroku:
C:\Program Files (x86)\ruby-1.9.2\bin\ruby.exe: No such file or directory -- /cygdrive/c/Program Files (x86)/Heroku/bin/heroku (LoadError)
What is this thing about ruby? I have no idea what I should be doing - developing in windows is a real pain, can anyone provide any insight into how I might solve this problem?
Appreciate any help. Thanks!
I ran in the same problem using Rails and Heroku on Windows.
It seems that the Toolbelt is not supported under Cygwin. Moreover colors are not always rendered in the right way (for example, I did not manage to render heroku logs colors, even after using ansicon -i).
I also considered using the CMD Prompt augmented with GOW but that means you have to append ".bat" to every command, and colors are still a problem.
I ended up using the Git Bash shell that is included with the RailsInstaller package.
It recongnizes all paths to relevant files, it has all the shell commands you need, and every color seems to be rendered correctly (e.g. rails logs, cucumber and rspec tests, heroku logs, etc.).
You've probably solved your issue a long time ago but I just wanted to add the steps I went through as I had the same issue on Windows with Cygwin.
Firstly always try to do an update of your Cygwin installation especially when you see an error similar to the one you've posted (I had the same error):
/ruby: No such file or directoryin/heroku: line 4: /cygdrive/d/Development/Heroku/ruby-1.9.2/bin
So I updated Cygwin and made sure to select all necessary ruby packages/interpreters etc, but this still didn't solve the problem as I kept getting the same error message.
Then I followed the steps outlined in Running the Heroku Command-Line Client Under Cygwin:
(1) Download RubyGem 1.9.3 from
http://rubyforge.org/frs/download.php/76072/rubygems-1.8.24.zip
(2) Then run the following -
$ unzip rubygems-1.8.24.zip
$ cd rubygems-1.8.24/rubygems-1.8.24
$ ruby setup.rb install
$ gem update --system
$ gem install heroku
(3) Open a new shell window and verify the version -
$ heroku version
heroku-gem/2.28.10 (i386-cygwin) ruby/1.8.7
This solved my problem and I can now run heroku commands from the Cygwin shell on Windows.
For me #Azkuma's answer only got me half the way. What worked for me:
1) Download and extract RubyGem zip: https://rubygems.org/pages/download
2) Set aliases to gem and heroku
alias gem='C:/ruby/bin/gem'
alias heroku='"C:/Program Files (x86)/Heroku/bin/heroku.bat"'
3) install as above
ruby setup.rb install
gem update --system
gem install heroku
4) login to heroku
heroku login
I found simply setting an alias worked for me.
alias heroku=c:/Program\\\ Files\\\ \\\(x86\\\)/Heroku/bin/heroku.bat
Then I can just use the heroku command directly with Cygwin.
The only thing I have a problem with is heroku login (and by extension, git push heroku master) whereby I'm prompted to use cmd.exe. For that part, I just open my Git Bash window from within the relevant folder, login and push from there.